Background
Intra-amniotic infection or inflammation is common in early preterm birth and associated with substantial neonatal lung morbidity owing to fetal exposure to proinflammatory cytokines and infectious organisms. Amniotic fluid interleukin 8, a proinflammatory cytokine, was previously correlated with the development of neonatal bronchopulmonary dysplasia, but whether amniotic fluid cytokines or placental pathology more accurately predicts neonatal lung pathology and morbidity is unknown. We have used a pregnant nonhuman primate model of group B Streptococcus infection to study the pathogenesis of intra-amniotic infection, bacterial invasion of the amniotic cavity and fetus, and microbial-host interactions. In this nonhuman primate model, we have studied the pathogenesis of group B Streptococcus strains with differing potential for virulence, which has resulted in a spectrum of intra-amniotic infection and fetal lung injury that affords the opportunity to study the inflammatory predictors of fetal lung pathology and injury.
Objective
This study aimed to determine whether fetal lung injury is best predicted by placental histopathology or the cytokine response in amniotic fluid or maternal plasma.
Study Design
Chronically catheterized pregnant monkeys ( Macaca nemestrina, pigtail macaque) at 116 to 125 days gestation (term at 172 days) received a choriodecidual inoculation of saline (n=5), weakly hemolytic group B Streptococcus strain (n=5, low virulence), or hyperhemolytic group B Streptococcus strain (n=5, high virulence). Adverse pregnancy outcomes were defined as either preterm labor, microbial invasion of the amniotic cavity, or development of the fetal inflammatory response syndrome. Amniotic fluid and maternal and fetal plasma samples were collected after inoculation, and proinflammatory cytokines (tumor necrosis factor alpha, interleukin beta, interleukin 6, interleukin 8) were measured by a multiplex assay. Cesarean delivery was performed at the time of preterm labor or within 1 week of inoculation. Fetal necropsy was performed at the time of delivery. Placental pathology was scored in a blinded fashion by a pediatric pathologist, and fetal lung injury was determined by a semiquantitative score from histopathology evaluating inflammatory infiltrate, necrosis, tissue thickening, or collapse scored by a veterinary pathologist.
Results
The principal findings in our study are as follows: (1) adverse pregnancy outcomes occurred more frequently in animals receiving hyperhemolytic group B Streptococcus (80% with preterm labor, 80% with fetal inflammatory response syndrome) than in animals receiving weakly hemolytic group B Streptococcus (40% with preterm labor, 20% with fetal inflammatory response syndrome) and in controls (0% preterm labor, 0% fetal inflammatory response syndrome); (2) despite differences in the rate of adverse pregnancy outcomes and fetal inflammatory response syndrome, fetal lung injury scores were similar between animals receiving the weakly hemolytic group B Streptococcus strains and animals receiving the hyperhemolytic group B Streptococcus strains; (3) fetal lung injury score was significantly correlated with peak amniotic fluid cytokines interleukin 6 and interleukin 8 but not tumor necrosis factor alpha or interleukin 1 beta; and (4) fetal lung scores were poorly correlated with maternal and fetal plasma cytokine levels and placental pathology.
Conclusion
Amniotic fluid interleukin 6 and interleukin 8 levels were superior predictors of fetal lung injury than placental histopathology or maternal plasma cytokines. This evidence supports a role for amniocentesis in the prediction of neonatal lung morbidity owing to intra-amniotic infection, which cannot be provided by cytokine analysis of maternal plasma or placental histopathology.
Introduction
Intra-amniotic infection or inflammation (IAI) is defined as intrauterine infection, inflammation, or both and is commonly associated with infection of the placenta and amniotic membranes. It occurs in only 3% to 5% of term pregnancies but greater than 50% of preterm deliveries before 26 weeks’ gestation. , IAI is traditionally associated with ascending infection from vaginal bacteria but can be seen with “sterile inflammation” in the absence of microbial infection. During IAI, there is production of proinflammatory cytokines, such as tumor necrosis factor alpha (TNF-α), interleukin 1 beta (IL-1β), IL-6, and IL-8, and infiltration of neutrophils in the placenta, amniotic membranes, and amniotic fluid (AF). , The combination of direct pathogen effects and exposure to proinflammatory cytokines and a cytotoxic inflammatory infiltrate result in the elevated maternal risk of preterm labor and preterm premature rupture of membranes (PPROM) and an increased risk of neonatal bronchopulmonary dysplasia, necrotizing enterocolitis, intraventricular hemorrhage, cerebral palsy, sepsis, and neonatal demise. Although prematurity is associated with neonatal morbidity, the addition of IAI has conferred an increased risk of neonatal complications in previous studies.
Why was this study conducted?
This study was conducted to determine whether specific features of placental pathology or maternal, fetal, or amniotic fluid (AF) cytokine levels best correlate with fetal lung injury.
Key findings
In this nonhuman primate model, fetal lung injury was significantly correlated with peak AF cytokines interleukin (IL)-6 and IL-8 but not with maternal plasma cytokines or placental pathologic features.
What does this add to what is known?
The findings support an expanded role for amniocentesis to evaluate cytokine response for the clinical prediction of fetuses at risk of adverse neonatal respiratory outcomes.
IAI is currently diagnosed and treated on the basis of clinical criteria; however, the severity of clinical symptoms and neonatal outcomes is not well correlated with placental histology (chorioamnionitis) or maternal, fetal, or intra-amniotic cytokine levels. , The effect of IAI on fetal lung injury, in particular, remains controversial despite bronchopulmonary dysplasia being the most common adverse outcome in severely preterm infants; heterogeneity across multiple potential variables makes analyzing this association complicated, including differences in pathogen exposures, duration of exposure before delivery, maternal and fetal immune responses, and postnatal exposures (eg, mechanical ventilation). Previous studies have suggested that higher neonatal IL-6 levels diagnostic of the fetal inflammatory response syndrome (FIRS; fetal plasma IL-6 of >11 pg/mL) are associated with an increased risk of neonatal adverse outcomes. , , However, whether FIRS is a superior predictor of fetal lung injury than placental histopathology or AF cytokines is currently unclear. ,
Because of the heterogeneity in clinical studies of IAI, animal models have been used to gain insight into the pathophysiology of the disease and mechanism of fetal injury. Findings from these models include correlations between placental and fetal neurologic inflammations and cerebrovascular hemodynamics in a rat model of lipopolysaccharide (LPS)-induced inflammation , ; fetal gut and neurologic inflammations in an LPS-induced model of IAI in sheep ; and fetal lung, kidney, and gut injuries associated with placental inflammation in a murine model. , These animal models demonstrate fetal injury associated with the concentration of inflammatory cytokines and chemokines in the AF and placenta, including TNF-α, interferon gamma, IL-1β, IL-4, IL-6, IL-10, and C-X-C-motif chemokine ligand-1 (CXCL1); an elevation in these cytokines can lead to recruitment of inflammatory cellular infiltrates of neutrophils, monocytes, and T cells in fetal tissues. Notably, these models typically rely on systemic or intra-amniotic inoculation of inflammatory mediators resulting in an overwhelming fetal sepsis and rapid preterm labor rather than the more typical subclinical IAI with a longer clinical course seen in human pregnancies. It has previously been demonstrated that subclinical inflammation can have a marked effect on fetal outcomes , and that key differences in fetal outcomes occur with exposure to intrauterine and systemic inflammations, suggesting the need for alternative animal models.
We have refined a chronically catheterized nonhuman primate (NHP) model to investigate the pathogenesis of IAI, preterm birth therapeutics, and the impact of exposure to different virulence factors associated with group B Streptococcus (GBS) strains. , We previously found that a choriodecidual inoculation of a weakly hemolytic GBS (WH GBS; eg, COH-1 ) strain can induce fetal lung injury without microbial invasion of the amniotic cavity owing to AF cytokine secretion by placental tissues. , Furthermore, GBS strains with a higher virulence potential, such as highly hemolytic strains (HH GBS; eg, GBSΔ covR ), are associated with preterm birth and adverse outcomes, such as microbial invasion of the amniotic cavity and fetal bacteremia. The use of GBS strains with lower and higher virulence has resulted in a spectrum of IAI severity in our NHP model that affords the opportunity for studying placental histopathologic and cytokine predictors of fetal lung injury. Our study objective was to determine whether placental histopathology or AF, maternal, or fetal cytokines were superior predictors of neonatal lung pathology at the time of birth.
Methods
Animals and study design
A description of the chronically catheterized pregnant NHP model of choriodecidual GBS infection, including details on animal care, analgesia, surgery, fetal euthanasia, necropsy, and fetal lung pathology and scoring, has been previously described. , Briefly, 15 chronically catheterized pregnant NHPs ( Macaca nemestrina ; pigtail macaque) at 118 to 125 days of gestation (term at 172 days) received 1 of the following 3 experimental treatments: (1) choriodecidual saline infusions (saline control group, n=5) ; (2) choriodecidual infusion with a WH GBS (COH-1 strain) 1×10 6 colony-forming units (CFU, n=5); (3) choriodecidual infusion with HH GBS (GBSΔ covR strain) ranging from 1×10 8 to 3×10 8 CFU (n=5) ( Table 1 ). Maternal blood, fetal blood, and AF samples were collected before exposure; at 0, 6, 12, 18, and 24 hours after exposure; daily after infection; and at the time of delivery. A cesarean delivery was performed at the experimental endpoint, which was either at the time of preterm labor, 4 days after bacterial inoculation, or 7 days after saline infusion. Labor was defined as progressive uterine activity associated with cervical effacement and dilatation. Intra-amniotic pressure was continuously recorded and digitized with a PowerLab system (AD Instruments, Colorado Springs, CO). Cervical examinations were performed on all animals the day before inoculation and then daily in the GBS group or every other day in the saline group until cesarean delivery. At the time of cesarean delivery, the delivery of the fetus was performed with tissue collection from the placenta and fetal necropsy. This study was conducted in strict accordance with the recommendations in the “Guide for the care and use of laboratory animals” of the National Research Council and the “Weatherall report on the use of NHPs in research.” The protocol was approved by the Institutional Animal Care and Use Committee of the University of Washington (permit number, 4165-01). All surgeries were performed while the animals were under general anesthesia, and all efforts were made to minimize suffering.
| Animal | Fetal sex | Gestational age at delivery (d) |
|---|---|---|
| Saline 1 | Female | 143 |
| Saline 2 | Male | 142 |
| Saline 3 | Female | 139 |
| Saline 4 | Male | 143 |
| Saline 5 | Male | 145 |
| WH GBS 1 | Female | 143 |
| WH GBS 2 | Male | 134 |
| WH GBS 3 | Male | 139 |
| WH GBS 4 | Female | 141 |
| WH GBS 5 | Male | 140 |
| HH GBS 1 | Male | 130 |
| HH GBS 2 | Male | 129 |
| HH GBS 3 | Female | 131 |
| HH GBS 4 | Male | 135 |
| HH GBS 5 | Female | 135 |
Detection of microbial invasion
Microbial invasion of the amniotic cavity was determined by plating AF and detection of GBS as previously described. In brief, AF (200 μL) from each sample time interval was serially diluted; 10-fold dilutions were plated on tryptic soy agar (TSA, Difco Laboratories Inc, Detroit, MI), incubated overnight at 37°C, 5% CO 2 and enumerated to determine bacterial invasion. To confirm the specific GBS strain, we took advantage of the orange color of the HH GBSΔ covR strain (hyperpigmented), whereas wild-type COH-1 are white in color. In addition, Δ covR strains of GBS are spectinomycin-resistant because the gene covR was replaced with a gene conferring spectinomycin resistance in these strains. , To confirm that the GBS strains recovered from infected NHP were the correct strains, a few colonies obtained from each sampled tissue and fluid per experiment were patched on a selective medium (ie, TSA-containing spectinomycin), and the level of CAMP factor activity was tested on sheep blood agar plates with the inoculum strain included in parallel. In addition, CHROMagar StrepB (DRG International, Inc, Springfield, NJ) was used as needed to confirm the presence of GBS bacteria.
Pathology and histology
Slides were made from formalin-fixed and paraffin-embedded sections of the placenta and fetal lungs and stained by hematoxylin and eosin. Lung slides were examined by a board-certified veterinary pathologist who was blinded to the group assignments. Lung histologic sections were evaluated and scored using a semiquantitative scale, as reported previously. Components were scored on a scale ranging from 0 to 4 (where 0 is normal) for inflammatory cells, necrosis, and inflammation, including tissue thickening, collapse, or other injuries. The lung components scored were (1) vascular or perivascular regions, (2) bronchial or peribronchial regions, (3) alveolar walls, and (4) trichrome stain intensity of the lung. Mononuclear inflammatory cells and neutrophils (identified using Leder staining) were counted within alveolar spaces in 5 random ×40 magnification fields. These parameters were incorporated into a severity score. Placenta histopathologic sections were examined by a board-certified pathologist who was blinded to group assignments. Histologic chorioamnionitis was defined by a maternal or fetal inflammatory response as determined by the Redline criteria. Maternal stage 1 describes acute subchorionitis, stage 2 is acute chorioamnionitis, and stage 3 is necrotizing chorioamnionitis. Maternal grade 1 reflects mild-moderate neutrophil infiltration, and grade 2 is severe confluent neutrophil inflammation. Fetal stage 1 is consistent with chorionic vasculitis, stage 2 is umbilical arteritis, and stage 3 is necrotizing funisitis. Fetal grade 1 describes mild-moderate neutrophil infiltration, and grade 2 is severe, confluent neutrophil infiltration. Sections without evidence of inflammation were considered stage 0 and grade 0.
Immunohistochemistry
Immunohistochemistry (IHC) staining was performed on placental tissues using mouse antihuman cluster of differentiation 8 (CD8) (1:100 dilution, Clone RPA-T8, BD number 557084, BD Biosciences, San Jose, CA) antibody or mouse immunoglobulin G (Abcam number 98952, Abcam, Cambridge, United Kingdom) isotype controls per previously described protocol. Briefly, slides were prepared from formalin-fixed and paraffin-embedded specimens of placental tissue. Dewaxing protocol was performed, and 10% normal goat serum was used for blocking. Staining was performed using BOND-automated IHC stainer (Leica, Wetzlar, Germany). Immunohistochemical cell staining was quantified using computer-assisted morphology (Image Pro, Miami, FL). The total number of CD8+ cells was determined from representative 5 high-powered fields on each specimen.
Cytokine analysis
Cytokine analysis was performed as described previously. Maternal and fetal blood and AF samples were obtained at various time intervals after inoculation, and centrifugation was performed. Sample supernatants were frozen at −80°C immediately after harvesting for later analysis. Cytokine concentrations of IL-1β, IL-6, IL-8, and TNF-α were determined using Luminex assay using commercial kits (Millipore, Burlington, MA). Typically, the coefficient of variation for Luminex assays is between 15% and 20%. FIRS was defined previously as fetal plasma IL-6 greater than 11 pg/mL.
Statistical analysis
Statistical analysis was performed using the Fisher exact test and chi-square for categorical variables, the Student t test on log-transformed data, analysis of variance to compare more than 2 groups with the Tukey honestly significant difference test to correct for multiple comparisons, log-rank test for survival curves, mixed-methods modeling for cytokine time course, and Kruskal-Wallis testing for ordinal data. All analyses were performed using Stata 14 software (StataCorp, College Station, TX). Statistical significance was defined as a P value of <.05. Because of the necessarily limited number of subjects in this NHP study , P values were noted in the text if <.1.
Results
Hyperhemolytic group B Streptococcus is associated with increased adverse clinical outcomes compared with weakly hemolytic group B Streptococcus
Previous work by our group and others has demonstrated that inoculation with GBS can introduce pathologic changes in fetal lung, cardiac, and brain tissues even in the absence of substantial intraamniotic infection, mimicking findings in neonates. , , , Because of these findings, the mechanism of how fetal lung injury develops during IAI has been difficult to characterize in part because of differences in pathogen type, which has led to difficulty in predicting the degree of neonatal lung injury in a fetus exposed to IAI. In this study, we used a subclinical model of IAI in NHPs to test how differences in GBS strain virulence affect fetal lung injury. Animals were inoculated within the choriodecidual space with low virulence WH GBS (GBS COH-1 strain), high virulence HH GBS (GBSΔ covR strain), or saline (controls), and clinical data are shown in Table 2 . Saline controls did not undergo preterm labor. Animals infected with HH GBS underwent preterm labor at an earlier time after inoculation (average day 2 vs day 5 after inoculation; P =.006) and at a higher rate (4 of 5 [80%] [4 of 5] vs 40% [2 of 5]; P =.009) than animals exposed to WH GBS. HH GBS was more likely to be recovered from the AF of infected animals (60% [3 of 5] vs 0%) than WH GBS. These results show that HH GBS is associated with increased adverse pregnancy outcomes compared with WH GBS.
| Characteristics | Saline (n=5) | Weakly hemolytic GBS (n=5) | Hyperhemolytic GBS (n=5) | P value |
|---|---|---|---|---|
| Fetal female sex | 2 (40) | 2 (40) | 2 (40) | NS |
| Fetal weight (g) | 314 (±70) | 309 (±19) | 272 (±64) | NS |
| Gestational age at delivery (d) | 142.4 (±2.2) | 139.4 (±3.4) | 132 (±2.8) | <.001 a |
| Preterm labor | 0 (0) | 2 (40) | 4 (80) | .009 b |
| Time from inoculation to labor (d) | — | 5 (4–7) | 2 (1–4) | .006 b |
| Microbial invasion of the amniotic cavity | 0 (0) | 0 (0) | 3 (60) | .1 |
| FIRS (fetal plasma concentrations of IL-6 >11 pg/mL) | 0 (0) | 1 (20) | 4 (80) | .03 c |
| Fetal lung score (scale, 0–5) | 0 (0–2) | 3 (0–4) | 1 (0–4) | .1 |
Cytokine differences between weakly and hyperhemolytic group B Streptococcus strains
To determine whether inflammatory factors varied by pathogen virulence in our controlled model, we obtained maternal, fetal, and AF cytokines at regular time intervals after inoculation. WH or HH GBS strains were not associated with a significant difference in levels of AF cytokines TNF-α and IL-1β ( Figure 1 , A and B). Both WH GBS– and HH GBS–exposed animals demonstrated similar levels of increased AF cytokine IL-8 compared with saline controls (WH GBS, 16.4 ng/mL [±11.7]; HH GBS, 15.6 ng/mL [±9.4]; saline, 1.4 ng/mL [±0.6]; P =.03) ( Figure 1 , C). Animals exposed to WH GBS had significantly higher peak levels of AF cytokine IL-6 than animals exposed to HH GBS (WH GBS, 81.1 ng/mL [±42.2]; HH GBS, 23.3 ng/mL [±3.9]; saline, 10.7 ng/mL [±4.9]; P =.002) ( Figure 1 , D). Maternal plasma cytokines were not found to be significantly different between WH and HH GBS strains and were similar to control animals, confirming a lack of maternal plasma cytokine elevation in our model, consistent with subclinical infection ( Figure 1 , E–H). There were significant increases in peak fetal inflammatory cytokines TNF-α ( P <.001) and IL-8 ( P =.03) but not in IL-1β and IL-6 in HH GBS infection compared with WH GBS infection ( Figure 1 , I–L).
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
